asia pacific research initiative for sustainable energy ......3 no major findings. hx design apv...
TRANSCRIPT
-
Asia Pacific Research Initiative
for Sustainable Energy Systems
2011 (APRISES11)
Office of Naval Research
Grant Award Number N0014-12-1-0496
OTEC Heat Exchanger Development and Testing
Task 4.1
Prepared For
Hawaii Natural Energy Institute
1680 East West Road, POST 109
Honolulu, HI, 96822
Prepared By
MAKAI OCEAN ENGINEERING
PO Box 1206, Kailua, Hawaii 96734
December 2014
-
1
OTEC HX TESTING PROGRAM
2014 ANNUAL REPORT
Prepared For
HAWAII NATURAL ENERGY INSTITUTE
RICK ROCHELEAU
1680 East West Road, POST 109
Honolulu, HI, 96822
USA
Prepared By
MAKAI OCEAN ENGINEERING
PO Box 1206, Kailua, Hawaii 96734
December 2014
-
2
1. EXECUTIVE SUMMARY
Makai Ocean Engineering, Inc.’s Ocean Thermal Energy Conversion (OTEC) Heat
Exchanger (HX) Testing program at the Ocean Energy Research Center (OERC) is focused on
developing an economically viable OTEC power plant. Heat exchangers are one of the most
expensive components in an OTEC power plant; proper heat exchanger selection is crucial to the
economic viability of OTEC. Heat exchanger development must balance size, cost, and
performance. To meet this goal, the OTEC HX Testing Program is divided into three areas: HX
Performance Testing, HX Design Development, and Corrosion Testing. This report reviews
Makai’s activities and summarizes the completion of the Milestones under APRISES11,
conducted from March 2013 to December 2014.
Major accomplishments in this period include:
HX Performance Testing Facility
Facility used to complete performance testing of APV Titanium Plate-Frame Heat
Exchanger, APV Cross Flow Heat Exchanger, and Lockheed Horizontal Heat Exchanger.
Modifications completed in preparation for turbine installation and testing.
Upgraded instrumentation and instrument wiring on the facility.
New revision to the control software.
Completion of 100-kW Testing Station.
HX Design Development
APV Titanium Plate-Frame Heat Exchanger was designed, fabricated, installed, and
performance tested.
APV Cross-Flow Heat Exchanger was designed, fabricated, installed, and performance
tested.
Lockheed Martin Horizontal Evaporator was designed, fabricated, installed, and
performance tested.
Corrosion Testing
Removal and analysis of 4-year hollow extrusion corrosion samples.
Ongoing testing of pit mitigation treatments.
Completed testing of alternative treatments to prevent biofouling.
Completed coatings testing.
Installed Lockheed Martin representative heat exchanger samples and plate samples for
the multi-column imaging rack (MCIR) – included warm seawater pre-treatment for
several samples.
Major findings in this period include:
HX Testing Facility
-
3
No major findings.
HX Design
APV Titanium Plate-Frame Heat Exchanger had high overall heat transfer coefficient
and very high seawater side pressure drop.
APV Cross Flow Evaporator had average overall heat transfer coefficient and very
low seawater side pressure drop.
Corrosion Testing
Hypochlorination is the most effective biofoulant control.
Siloxel and Alodine coatings provide protection but coating uniformity was unreliable
and therefore, corrosion protection was limited.
Steel is no longer being pursued as a potential material for a condenser.
Pit mitigation treatments should be performed based on the sample’s open circuit
potential (OCP), not set intervals.
-
4
2. HX PERFORMANCE TESTING FACILITY
As of October 2014, the HX Performance Testing Facility is at capacity with six heat
exchangers installed. On the condenser side, the Lockheed Martin Shell and Tube Heat
Exchanger, Lockheed Martin Enhanced Tube Heat Exchanger, and the APV Titanium Plate-
Frame Heat Exchanger are installed. On the evaporator side, CHART’s Brazed Aluminum Heat
Exchanger, Lockheed Martin’s Horizontal Shell and Tube Heat Exchanger, and APV’s Cross-
Flow Heat Exchanger are installed.
Facility instrumentation was upgraded to include new ultra-sonic seawater flow sensors,
an additional liquid ammonia Coriolis flow meter in the separator-receiver line, a level sensor in
the separator-receiver line, and a pair of precise temperature probes for during heat exchanger
testing. Instrument wiring was overhauled in June 2014 with new conduit, junction boxes, multi-
conductor cabling, and standardized terminal blocks. A positive pressure system was also
installed to supply dry compressed air to aid in keeping instrumentation dry. Instrument
calibration was performed in August and September 2014. Pressure, temperature, and flow
sensors were calibrated in preparation for upcoming HX performance testing.
-
5
3. HX DESIGN AND DEVELOPMENT
Makai completed designs for two APV titanium plate-frame style heat exchangers. Both
have been tested at the HX Performance Testing Facility. A Lockheed Martin Horizontal Shell
and Tube Heat Exchanger was also tested at the HX Performance Testing Facility during this
period.
The key parameters of heat exchanger performance that directly affect OTEC plant
design are overall heat transfer coefficient, seawater-side pressure drop, and ammonia-side
pressure drop. The overall heat transfer coefficient (U) is a measure of the heat exchanger’s
efficiency. Heat exchangers with higher U values require less surface area to transfer a given
duty. This is important for OTEC because more efficient heat exchangers require less space
which equates to big savings on the cost of the remoras. The ammonia-side and seawater-side
convective heat transfer coefficients were also calculated using the definition of the overall heat
transfer coefficient as a function of the convective and conductive heat transfer coefficients. h1
and h2 are the ammonia-side and water-side convective heat transfer coefficients, k is the
conductivity of aluminum and dx is the wall thickness of the titanium plate extrusion.
In order to determine h1 and h2, the water-side heat transfer coefficient was assumed to be
constant for each water flow rate set point and the ammonia-side heat transfer coefficient was
assumed to be constant for each duty set point. The method of least squares was then used to
determine a single heat transfer coefficient for each set point.
Seawater-side pressure drop affects the amount of OTEC-generated power that must be
used to supply seawater pumps on an OTEC plant. High pressure drops require large amounts of
pumping power, which reduces the net-power output from the OTEC plant. Ammonia-side
pressure drop is not expected to be a significant factor.
Ammonia-side operating pressure and approach temperature also provide valuable insight
to heat exchanger performance and further our understanding of design decisions. Ammonia-
side operating pressure is related to power production because the power generated in an OTEC
plant is a function of the pressure drop across the turbine and the ammonia flow rate. A lower
pressure on the condenser side and higher pressure on the evaporator side is generally better for
OTEC as this increases the pressure drop across the turbine and creates higher power output.
Approach temperature is the difference in temperature of the two fluids at the heat exchanger
outlet. Small approach temperatures indicate the amount of heat transferred toward the end of
the heat exchanger is greatly diminished due to small temperature differences between the two
fluids.
-
6
3.1. APV TITANIUM PLATE-FRAME HEAT EXCHANGER
The APV titanium plate-frame heat exchanger was designed and tested as a condenser.
APV and Makai worked together to customize the design for OTEC conditions. In a plate-frame
heat exchanger, fluid flows in channels between a series of stacked plates. Gaskets are used to
seal the plates. A rigid steel frame is used to hold the plates together and maintain gasket
compression. Each plate has a port in two opposing corners to allow either hot or cold fluid to
enter. The use of larger plates in the APV condenser (plates are larger than plates in existing
plate-frame heat exchangers that Makai has researched) allows for fewer manifold pipes in a full-
scale OTEC power plant. Titanium was selected as a low-risk solution for a condenser. Pitting,
which has been observed in corrosion testing of aluminum alloys, remains a concern regarding
the longevity of aluminum condensers, especially in cold seawater. Although the initial cost of a
titanium heat exchanger is more than a comparable aluminum heat exchanger, the aluminum heat
exchanger may not last the life of an OTEC plant and required additional maintenance and
replacement costs.
Table 3-1. APV Titanium Plate-Frame Heat Exchanger Specifications
Plate Properties
Material Titanium ASTM B265 Grade1
Number of Plates 97 total
Plate Thickness 0.5 mm
Total heat transfer area 214.89 m2 total
Frame Properties
Material Carbon Steel
Frame Height 3.74 m
Frame Width 1.28 m
Frame Length 3.30 m
Max Number of Plates Pairs 48 + 1 single
The APV condenser arrived in the OERC on October 14, 2013 and was installed over a
ten-day span. Performance testing was conducted on October 28, 2013. A total of 19 operating
points were tested (Table 3-2). Each operating point was held at steady state for at least ten
minutes. Steady state was defined as being within +/- 3% for duty, +/- 2% for seawater flow
rate, +/- 0.1°C for seawater inlet temperature, +/-0.2°C for seawater temperature difference, +/-
2% for LMTD, +/- 2% for ammonia flow rate, +/- 2% for quality, and +/- 1 kPa for condenser
pressure.
-
7
Figure 3-1. APV Titanium plate-frame heat exchanger installed on the OERC HX
Performance Testing Facility.
Table 3-2. APV Condenser Performance Testing Points
Duty [kW]
1000 1500 2000 2500 3000
Cold
Wate
r F
low
[gp
m]
1000 X X X
1500 X X X X
2000 X X X X
2500 X X X X
2900 X X X X
-
8
3.1.1. Overall heat transfer coefficient
U is plotted below in two ways – with lines of constant seawater flow and with lines of
constant duty. U is mostly dependent on water velocity and less dependent on duty, indicating
the heat exchanger is limited by the water-side heat transfer rate.
Figure 3-2. Condenser overall heat transfer coefficient
2.00
2.50
3.00
3.50
4.00
4.50
500 1000 1500 2000 2500 3000 3500
U [
kW/m
2-K
]
Q NH3 [kW]
Overall Heat Transfer Coefficient vs Duty
1000 gpm 1500 gpm 2000 gpm 2500 gpm 2900 gpm
2.00
2.50
3.00
3.50
4.00
4.50
500 1000 1500 2000 2500 3000 3500
U [
kW/m
2-C
]
Cold Seawater Flow Rate [gpm]
Overall Heat Transfer Coefficient vs Flow
1000 kW 1500 kW 2000 kW 2500 kW 3000 kW
-
9
3.1.2. Convective heat transfer coefficient
In the tested range, the overall heat transfer coefficient was mainly influenced by the
seawater convective heat transfer coefficient which varied from 4 to 14 kW/m2-C whereas the
ammonia convective heat transfer coefficient showed little change (6.5 to 7.5 kW/m2-C).
Figure 3-3. Ammonia and seawater convective heat transfer coefficients fitting the equation
1/U = 1/hsw + dxw/k + 1/hNH3 solved for using the least–squares method.
3.1.3. Seawater-side pressure drop
The seawater-side pressure drop is independent of duty and has a power-law relationship
to water velocity. The power-law relationship has an exponent of ~2.7. Pressure drop increases
exponentially with increased water flow, opposite to the trend of U value, which shows
diminishing increases with increased flow. These opposing trends imply there is an optimum
water velocity that balances increased U value with increased pressure drop across the
condenser.
y = 0.0049x - 1.1644 R² = 0.99
y = -3E-07x2 + 0.0005x + 7.3277 R² = 0.8951
0
2
4
6
8
10
12
14
16
0 500 1000 1500 2000 2500 3000 3500
Co
nve
ctiv
e H
eat
Tran
sfer
Co
effi
cien
t [k
W/m
2-K
]
Seawater Flow [gpm], Duty [kW]
Convective Coefficients
Seawater
Ammonia
-
10
Figure 3-4. Seawater-side pressure drop
3.1.4. Ammonia-side pressure drop
The magnitude of the pressure drop on the ammonia side is much smaller than the
pressure drop on the seawater side. Also, one of the ammonia pressure sensors may be faulty –
the dP across the condenser should never be negative.
Figure 3-5. Ammonia-side pressure drop
y = 5.69E-08x2.66 R² = .996
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0
500 1000 1500 2000 2500 3000 3500
SW d
P [
kPa]
Cold Seawater Flow Rate [gpm]
Seawater Pressure Drop
1000 kW 1500 kW 2000 kW 2500 kW 3000 kW
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
500 1000 1500 2000 2500 3000 3500
CIP
S -
CO
PS
[kP
a]
Q NH3 [kW]
Ammonia Pressure Drop
1000 gpm 1500 gpm 2000 gpm 2500 gpm 2900 gpm
-
11
3.1.5. Ammonia-side operating pressure
For a given duty, the condenser pressure decreases with increased seawater flow rate.
This means that the power output should be greater for higher water flow rates. However, the
increased power with increasing seawater flow will eventually be offset by the increased
pumping power at higher flows (due to the increase in seawater-side pressure drop).
Figure 3-6. Ammonia absolute pressure at condenser inlet
3.1.6. Approach Temperature
Ammonia duty is used in this calculation because the ammonia pressure sensors are
used to calculate the ammonia temperature at saturation and provide a more accurate temperature
measurement than the temperature sensors on the seawater side.
550
570
590
610
630
650
670
690
710
730
500 1000 1500 2000 2500 3000 3500
CIP
S [k
Pa]
Cold Seawater Flow Rate [gpm]
Ammonia Inlet Pressure
1000 kW 1500 kW 2000 kW 2500 kW 3000 kW
-
12
Figure 3-7. Approach temperature
The data shows a linear relationship between duty and approach temperature, with the
slope of the line dependent on the water velocity. The approach temperature increases with duty
because higher duty corresponds to higher ammonia operating pressure, and thus a higher
saturation temperature.
3.1.7. Difference between ammonia and seawater duty
Theoretically, if the condenser was perfectly insulated, the ammonia duty should be
identical to the seawater duty. However, during testing heat from the outside environment and
sensor accuracies can cause the two duties to be unequal. In general, the agreement between the
two duties was very good. At flow rates higher than 2500 gpm duties were within 5% (Figure
3-8). At lower duties and low flow rates, the discrepancy between the two measured values is
magnified, with the worst being 22% off.
0.0
0.5
1.0
1.5
2.0
2.5
500 1000 1500 2000 2500 3000 3500
CO
TS -
CW
TS2
[C
]
Q NH3 [kW]
Approach Temperature
1000 gpm 1500 gpm 2000 gpm 2500 gpm 2900 gpm
-
13
Figure 3-8. Difference between ammonia duty and seawater duty
-50
0
50
100
150
200
250
300
350
400
0 500 1000 1500 2000 2500 3000 3500
Q S
W -
Q N
H3
[kW
]
Cold Seawater Flow Rate [gpm]
Duty Difference
1000 kW 1500 kW 2000 kW 2500 kW 3000 kW
-5.0
0.0
5.0
10.0
15.0
20.0
25.0
30.0
0 500 1000 1500 2000 2500 3000 3500
Q S
W /
Q N
H3
- 1
[%
]
Cold Seawater Flow Rate [gpm]
Duty Difference
1000 kW 1500 kW 2000 kW 2500 kW 3000 kW
-
14
3.1.8. Design Discussion
The APV Titanium Plate and Frame Heat Exchanger had high overall heat transfer
coefficient but also high seawater pressure drop. The ammonia-side convection coefficient
limits the overall heat transfer coefficient, so it may be possible to achieve comparable
performance with lowered seawater pressure drop by lower resistance to flow on seawater
channels.
3.2. APV TITANIUM CROSS FLOW PLATE-FRAME HEAT EXCHANGER
The APV Cross Flow Evaporator is titanium plate-frame type heat exchanger. This heat
exchanger has a nominal thermal capacity of 2MW, like the previous five heat exchangers tested
at OERC. The plate size and type are identical to those within the previously tested APV
condenser. However, the shape of the corrugations are modified to optimize convection to
increase boiling.
Similar to the previously tested APV titanium plate-frame condenser, the APV cross flow
evaporator uses large plates which reduces the total number heat exchangers required, and thus
the number of manifold pipes and flange connections that will be required in a large OTEC
power plant. To further reduce manifold costs, Makai and APV designed a cross-flow
evaporator that eliminates the need for seawater manifolds, leading to a 50% reduction in
manifold costs. This new design required custom tooling to fabricate the plates. Seawater enters
through the side and passes through horizontally. Furthermore, the design of the heat exchanger
provides a more elegant mounting solution inside the Remora (the offshore superstructure which
houses all power cycle components, including heat exchangers, tanks, water pumps, etc.),
allowing for a reduction in the total size of the system.
Table 3-3. Evaporator Properties
Plate Properties
Material Titanium ASTM B265 Grade1
Number of Plates 149 total, 147 heat transfer
Plate Thickness 0.6 mm
Total heat transfer area 332.514 m2 total, 277.8 m
2 effective
Frame Properties
Material Carbon Steel
Frame Height 3.24 m
Frame Width 1.73 m
Frame Length 1.73 m
Max Number of Plates Pairs 74 + 1 single
-
15
The APV cross-flow evaporator was installed and tested in October 2014. A total of 21
operating points were tested (Table 3-4). Each operating point was held at steady state for at
least ten minutes. Steady state was defined as being within +/- 3% for duty, +/- 2% for seawater
flow rate, +/- 0.1°C for seawater inlet temperature, +/-0.1°C for seawater temperature difference,
and +/- 2% for LMTD.
Figure 3-9. APV Cross Flow Heat Exchanger installed at the OERC HX Performance Testing
Facility.
Table 3-4. APV Cross Flow Evaporator Test Points Matrix
Duty [kW]
1500 2000 2500 3000
Warm
Wate
r F
low
(gp
m)
1500 X X
2000 X X X
2500 X X X X
3000 X X X X
3500 X X X X
4000 X X X X
-
16
3.2.1. Overall heat transfer coefficient
U was mostly dependent on water velocity and less dependent on duty. This indicates that
the heat exchanger is limited by the water-side heat transfer rate.
The effect of quality on U was not explicitly tested but two test points collected at
different qualities illustrate U is lower for high qualities. The 4000 gpm – 2500 kW (green filled
circle) operating point and one of the 4000 gpm – 3000 kW (red “x”) operating points were
collected at high quality (82% and 94%, respectively) compared to the other points at ~ 60 %
quality. At 4000 gpm – 2500 kW, the U value at 82% quality is slightly lower than expected
from the trend at ~60% quality; the slope of the gray dashed line is not as steep as the solid gray
line. U value for the 94% quality point at 4000 gpm – 3000 kW is 2% lower than for the 60%
quality point at 4000 gpm – 3000 kW. Previous testing indicated U is mostly independent of
quality variations between 30-80% but decreases with increasing quality at qualities higher than
80%.
1.7
1.9
2.1
2.3
2.5
2.7
2.9
1000 1500 2000 2500 3000 3500
U [
kW/m
2/C
]
Q NH3 [kW]
Overall Heat Transfer Coefficient vs Duty
1500 gpm 2000 gpm 2500 gpm3000 gpm 3500 gpm 4000 gpm
-
17
Figure 3-10. Overall heat transfer coefficient
3.2.2. Convective heat transfer coefficients
In the tested range, the overall heat transfer coefficient was mainly influenced by the
seawater convective heat transfer coefficient which varied from 3.5 to 11 kW/m2-C whereas the
ammonia convective heat transfer coefficient showed little change (3.8 to 4.3 kW/m2-C).
Figure 3-11. Ammonia and seawater convective heat transfer coefficients fitting the equation
1/U = 1/hsw + dxw/k + 1/hNH3 solved for using the least–squares method.
1.7
1.9
2.1
2.3
2.5
2.7
2.9
1000 1500 2000 2500 3000 3500 4000 4500
U [
kW/m
2/C
]
WSW Flow [gpm]
Overall Heat Transfer Coefficient vs Flow
1500 kW 2000 kW 2500 kW 3000 kW
y = 0.003x - 1.1023 R² = 0.9986
y = 0.636ln(x) - 0.809 R² = 0.9924
3
4
5
6
7
8
9
10
11
1000 1500 2000 2500 3000 3500 4000 4500
Co
nve
ctiv
e H
eat
Tran
sfer
Co
effi
cien
t [
kW/m
2-K
]
Seawater Flow [gpm], Duty [kW]
Seawater and Ammonia Convective Coefficients
Seawater-Side
Ammonia-Side
-
18
3.2.3. Seawater-side pressure drop
The seawater-side pressure drop is independent of duty and has a power-law relationship
to flow rate with an exponent of ~1.9. However, with less than 10 kPa pressure drop at
4000 gpm, the pumping power requirement is quite low.
Figure 3-12. Seawater-side pressure drop
3.2.4. Ammonia-side pressure drop
Ammonia-side pressure drop is independent of seawater flow, increases with increasing
duty, and decreases with increasing quality because it is dependent on ammonia mass flow rate
(Figure 3-13). The black triangles are points collected at high quality (82% and 94%).
y = 1.4766E-06x1.8911 R² = .99953
0
2
4
6
8
10
12
0 500 1000 1500 2000 2500 3000 3500 4000 4500
SW d
P [
kPa]
WWFS [gpm]
dP H2O
-
19
Figure 3-13. Ammonia-side pressure drop. Triangle data points were collected at higher
quality than the other points. Since quality effects were not specifically tested, the high quality
points are considered outliers, shown for completeness but not included in analysis.
As measured, ammonia-side pressure drop is 5 times higher than the predicted but the as-
measured pressure drop also includes the effects of static head and momentum head along with
frictional head loss. By solving for the void fraction at different heights along the evaporator, the
static head and momentum head contribution to the total pressure drop can be calculated, leaving
the only the pressure drop due to frictional losses (Figure 3-14).
10
10.5
11
11.5
12
12.5
13
13.5
14
14.5
15
1000 1500 2000 2500 3000 3500
EIP
S -
EOP
S [k
Pa]
Q NH3 [kW]
Ammonia Pressure Drop
1500 gpm 2000 gpm 2500 gpm3000 gpm 3500 gpm 4000 gpm
y = 0.7687x2 - 3.0315x + 14.075 R² = 0.9607
10
10.5
11
11.5
12
12.5
13
13.5
14
14.5
15
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
NH
3 d
P [
kPa]
EIFS [kg/s]
NH3 dP
-
20
Figure 3-14. Ammonia-side frictional losses.
3.2.5. Ammonia-side operating pressure
For a given duty, the evaporator pressure increases with increasing seawater flow. This
means that gross power is greater for higher seawater flow rates; however, seawater pressure
drop increases exponentially with increasing seawater flow which increases pumping power and
ultimately reduces net power. As with the overall heat transfer coefficient, there is an optimum
point which maximizes the gain in power output with increased seawater flow while accounting
for the increase in pumping power required at higher seawater flow rates to produce the
maximum net power.
Figure 3-15. Ammonia absolute pressure at evaporator outlet increases with increasing
seawater flow rates.
-5
0
5
10
15
20
0 1 2 3 4 5
Pre
ssu
re D
rop
(kP
a)
Ammonia Flow Rate (kg/s)
Ammonia Pressure Drop
Total Pressure Drop Predicted Friction Losses Friction Losses
860
880
900
920
940
960
980
1000
1020
1000 1500 2000 2500 3000 3500 4000 4500
EOP
S [k
Pa]
WSW Flow [gpm]
Evaporator Outlet Pressure
1500 kW 2000 kW 2500 kW 3000 kW
-
21
3.2.6. Difference between ammonia and seawater duty
Theoretically, if the evaporator was perfectly insulated, the ammonia duty should be
identical to the seawater duty. However, environmental input and limitations in sensor accuracy
can cause the two duties to be unequal. The two duties were within 6% (Figure 3-16). Previous
testing showed greater duty discrepancy at lower flow rates and lower duties. Duty discrepancy
is greater at lower duties but contrary to previous testing, higher flow rates also resulted in higher
discrepancies for the same duty.
Figure 3-16. Difference between ammonia duty and seawater duty
3.2.7. Approach Temperature
The data shows a linear relationship between duty and approach temperature, with the
slope of the line dependent on the water velocity. The approach temperature increases with duty
because higher duty corresponds to higher ammonia operating pressure, and thus a higher
saturation temperature.
0
1
2
3
4
5
6
1000 1500 2000 2500 3000 3500
(Q N
H3
- Q
SW
)/Q
NH
3 *
10
0 [
[%]
Q NH3 [kW]
Duty Difference
1500 gpm 2000 gpm 2500 gpm
3000 gpm 3500 gpm 4000 gpm
-
22
Figure 3-17. Approach temperature
3.2.8. Design Discussion
Similar to the APV condenser, the ammonia-side convection coefficient limited the
overall heat transfer coefficient. For the cross-flow evaporator, seawater pressure drop was very
low so higher flow rates could be used. Design adjustments to enhance ammonia-side
convection could improve the evaporator performance. As tested, the APV Cross Flow
evaporator performance was average but likely need improvements for an OTEC heat exchanger.
3.3. LOCKHEED MARTIN HORIZONTAL EVAPORATOR
The Lockheed Martin Horizontal Evaporator was installed and tested in October 2014.
34 operating points were tested, some points were tested with variations in quality, evaporator
level, and pre-heating. Makai gained insight into the operation of a horizontal evaporator and
differences in performance of horizontally versus vertically oriented heat exchangers.
1.5
1.7
1.9
2.1
2.3
2.5
2.7
2.9
3.1
1000 1500 2000 2500 3000 3500
WW
TS2
- T
sat
@ E
OP
S [C
]
Q NH3 [kW]
Approach Temperature
1500 gpm 2000 gpm 2500 gpm3000 gpm 3500 gpm 4000 gpm
-
23
4. CORROSION TESTING
Makai Ocean Engineering, Inc. has been conducting corrosion testing in support of
OTEC heat exchanger design at OERC since 2009. Potential heat exchanger materials and
fabrication methods are tested in conditions representative of OTEC operations. Aluminum
alloys, which were selected for their cost-effectiveness and variety of fabrication options, have
been extensively tested. Samples tested at the OERC are exposed to seawater representative of
OTEC operating conditions (warm surface seawater (WSW), cold seawater (CWS) from ~674 m,
and deep seawater (DSW) from ~915 m).
Uniform general corrosion rates for aluminum alloys are acceptable but localized
corrosion (pitting and crevice corrosion), particularly in CSW, is of concern. Aluminum alloys
have performed well in WSW and show promise for use in an evaporator. In particular, Alloy
3003 has performed the best in WSW. Alloy performance has differed between the cold
seawater sourced from 674-m and 915-m. In 674-m CSW, Alloys 3003 and LA83I have
performed the best, but certain samples of each alloy had deep pits. In 915-m CSW, all alloys
have severe pits. Alloy 3003 has performed the best, but still had a ~ 0.5 mm deep pit after four
years of exposure.
After the first year of corrosion testing, Makai determined the established analytical
methods were inadequate for OTEC purposes and began to develop new techniques to evaluate
corrosion performance. Results from early tests also revealed poor performance of some
samples and has prompted Makai to investigate treatments that could improve corrosion
resistance. Makai developed instrumentation (daily, automated, high resolution imaging; in situ
ultrasonic pit detection; and laser profilometry) to quantify the extent and severity of pitting.
Makai has also been investigating treatments (curing in WSW, ozone flushing, and acid flushing)
and coatings to prevent and/or mitigate the development of localized corrosion. Finally, Makai
also evaluated biofouling controls.
4.1. BOX COUPONS
Box coupons have been tested since 2009. Four-year samples were removed in January
2014. All alloys are performing well in WSW, with little to no pitting. Pits are prevalent in
Alloy 1100 and 6063 samples exposed to 674-m CSW and 915-m seawater. Alloys 3003, 5052,
LA83I, and LA83P showed little pitting in 674-m CSW but all four alloys had significant pitting
in 915-m CSW. Samples moved from surface seawater to 674-m CSW after 40 days have begun
to show increasing pits.
Several changes have been made to the reporting of pitting data. The maximum pitting
depth is reported in place of the average of the maximum pit depths for all the samples of each
alloy in each water source. Since a single failure in a heat exchangers constitutes failure of the
heat exchanger, the maximum for all samples of the same alloy in each water source is now
reported.
-
24
To capture the extent of pitting, each scanned surface was divided into 0.5 x 0.5 mm
squares. The maximum pitted depth in each 0.5 x 0.5 mm square was recorded and the results
were plotted in histogram form for each sample. Samples of the same alloy in the same water
were also combined into one histogram for overall alloy performance per water source, over
time.
Corrosion product buildup can reduce heat exchanger performance. Since heat exchanger
performance is function of the available heat transfer surface and measured for the entire heat
exchanger, the average pitted area, as a percentage of total area, is reported. However, corrosion
product accumulates beyond the residual pitted area (up to several times the pitted area);
therefore, the actual heat transfer area that is affected by corrosion will be greater than the
average pitted area. The relationship between pitted area and corrosion buildup area varies
between the different alloys and warrants further investigation.
4.1.1. WSW
Table 4.1-1 summarizes the number of sample scanned per alloy, per exposure time. For
samples removed at the earlier exposure times, not all samples that were removed were scanned.
Table 4.1-1. Number of scanned WSW samples.
1yr 1.5yr 21mo 2yr 3 year 4 year
1100 - 2 - 4 2 4
3003 - 2 - 4 2 4
5052 - 3a - 4 2 4
6063 1 2 - 4 2 4
LA83I - 2 - 4 2 4
LA83P - 2 - 4 2 4
a 3 of 4 samples were scanned, scan quality was poor on 1 of the 3 scanned samples;
data from poor scan not shown in figures or used in analysis.
For the four-year samples, profilometer scans identified one pit each on two samples (one
of Alloy 1100 and one of Alloy LA83I, Figure 4-1) although the pits could not be visually
identified under a microscope. No pits were observed on the remaining 22 samples.
-
25
Figure 4-1. Maximum pit depth and average percent pitted area (per alloy, per exposure time)
for samples exposed to WSW. There are no scan data for 1-yr exposure samples of Alloys
1100, 3003, 5052, LA83I and LA83P.
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
1 yr 1.5 yr 2 yr 3 yr 4 yr
Max
imu
m P
it D
epth
[m
m]
Time Exposed
WSW - Maximum Pit Depth
1100 3003 5052 6063 LA83I LA83P
0.00001
0.0001
0.001
0.01
0.1
1 yr 1.5 yr 2 yr 3 yr 4 yr
Ave
rage
Per
cen
t P
itte
d A
rea
[%]
Time Exposed
WSW - Average Percent Pitted Area
1100 3003 5052 6063 LA83I LA83P
-
26
(a)
(b)
Figure 4-2. (a) Representative images and (b) weight loss results of samples exposed in WSW
for four years.
All alloys continue to perform well in WSW. The four-year samples had less pitting than
the three-year samples. Weight loss results also indicate the four-year samples were performing
better than the three-year samples. This trend will have to be evaluated when the five-year
samples are analyzed.
-
27
4.1.2. Cold Seawater (CSW)
For Alloys 1100 and 6063, pitting was severe. Both samples of Alloy 6063 had deep
extrusion defects which have been observed to be either pit initiation sites or sites more
susceptible to pitting. Both Alloy 1100 samples exhibited crevice corrosion. One sample even
had severe corrosion on the exterior surface, likely caused by leakage stemming from crevice
corrosion.
Table 4.1-2 summarizes the number of sample scanned per alloy, per exposure time. Not
all the 1-yr and 21-month samples that were removed were scanned.
Table 4.1-2. Number of scanned CSW samples/total removed samples.
1yr 1.5yr 21mo 2yr 3 year 4 year
1100 4 4 - 4 2 2
3003 - 4 - 4 1 2
5052 - 4 8 4 2 2
6063 2 4 9 4 2 2
LA83I - 4 - 4 2 2
LA83P - 4 1 4 2 2
One CSW rack was inspected at 21 months due to excessive leaking. Three samples of
LA83I were removed and visually inspected to have pits but were not scanned by the
profilometer. An additional three pitted samples of Alloy 6063 were removed after 21 months
but not scanned (12 samples removed total). Crevice corrosion induced leaks likely contributed
to further worsening of crevice corrosion and pit development. Since samples were visually
inspected and the worst samples were removed, the data are biased.
For the 4-yr samples, some discrepancies between the profilometer analysis and
microscope inspection were identified. Pits identified on LA83I-35 correspond to a relatively
large area where the sample surface appears gouged (Figure 4-3). Profilometer scans cannot
distinguish between scratches or other pre-existing surface defects and pits. Some variability is
expected due to surface imperfections. For both Alloy 5052 samples, profilometer scans
identified several shallow pits but visual inspection identified one pit on one Alloy 5052 sample
and no pits on the other sample. With Alloys 3003 and LA83P, pits identified on profilometer
scans were also not identified visually. Rescanning the samples did not resolve the discrepancy;
however, since the maximum pit depth for all four alloys was ~ 0.1 mm, it is more likely the pits
were not discernible under the microscope or some surface features biased the algorithm for pit
identification. Pitting was not severe in the four-year samples of Alloys 3003, 5052, LA83I, and
LA83P.
-
28
Figure 4-3. Pitted area identified in profilometer scan on LA83I-35. Under microscope
inspection, the area does not display characteristics of pitting, rather the surface appears
gouged or eroded.
-
29
Figure 4-4. Maximum pit depth and average percent pitted area (per alloy, per exposure time)
for samples exposed to CSW. There are no scan data for 1-yr exposure of Alloys 3003, 5052,
LA83I, and LA83P and 21-month exposure of Alloys 1100, 3003, and LA83I.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
1 yr 1.5 yr 21 mo 2 yr 3 yr 4 yr
Max
imu
m P
it D
epth
[m
m]
Time Exposed
CSW - Maximum Pit Depth
1100 3003 5052 6063 LA83I LA83P
0.00001
0.0001
0.001
0.01
0.1
1
10
1 yr 1.5 yr 21 mo 2 yr 3 yr 4 yr
Ave
rage
Per
cen
t P
itte
d A
rea
[%]
Time Exposed
CSW - Average Percent Pitted Area
1100 3003 5052 6063 LA83I LA83P
-
30
(a)
(b)
Figure 4-5. (a) Representative images and (b) weight loss results of samples exposed in CSW
for four years.
There are some anomalies in weight loss results for the 21 month and 2 year samples.
Alloys 5052 and 6063 had lower three- and four-year weight loss than 21-month and two-year
weight loss. 21-month the 2-year data were obtained from samples in different CSW racks but
severe crevice corrosion was noted in Alloy 5052 and 6063 samples for both exposure times.
-
31
4.1.3. Deep Seawater (DSW)
All alloys performed poorly in DSW. In addition to pitting corrosion, crevice corrosion
was also severe. Table 4.1-3 summarizes the number of sample scanned per alloy, per exposure
time. Although 1 year samples of all alloys were removed, only Alloy 1100 samples were
scanned.
Table 4.1-3. Number of scanned Deep Seawater ( DSW) samples.
1yr 1.5yr 21mo 2yr 3 year 4 year
1100 6 1 - 3 2 2
3003 - 1 - 4 2 2
5052 - 1 - 4 2 2
6063 - 1 - 4 2 2
LA83I - 1 - 4 2 2
LA83P - 1 - 4 2 2
Compared to other water sources, pitting and crevice corrosion was worst in DSW; nearly
all samples had pits and crevice corrosion. Several samples were so severely corroded that
corrosion product removal was incomplete; the actual maximum pit depth and pitted area is
higher than that recorded by the profilometer. Useful profilometer scans are limited to the flat
coupon surfaces, away from sides and edges; hence the extent of crevice corrosion is not entirely
captured by a scan and visual observations are required.
Both LA83P samples also had large pits on the sides of the samples which could not be
scanned by the profilometer.
Figure 4-6. One of several large pits on LA83P samples in DSW. These pits are located along
the sides of the sample and cannot be scanned by the profilometer.
-
32
Figure 4-7. Maximum pit depth and average percent pitted area (per alloy, per exposure time)
for samples exposed to DSW. Pits were prevalent on DSW samples. Only Alloy 1100 1-year
samples were scanned. No profilometer data is reported for the 1-year samples of the other
alloys. (No pits were observed, only two of Alloy 5052 samples had small blisters but no pit
craters). Many of the 4-year samples had pits on the edges and sides of the coupon that could
not be captured in the profilometer scans.
0
0.5
1
1.5
2
2.5
3
1 yr 1.5 yr 2 yr 3 yr 4 yr
Max
imu
m P
it D
epth
[m
m]
Time Exposed
DSW - Maximum Pit Depth
1100 3003 5052 6063 LA83I LA83P
0.00001
0.0001
0.001
0.01
0.1
1
10
100
1 yr 1.5 yr 2 yr 3 yr 4 yr
Ave
rage
Per
cen
t P
itte
d A
rea
[%]
Time Exposed
DSW - Average Percent Pitted Area
1100 3003 5052 6063 LA83I LA83P
-
33
(a)
(b)
Figure 4-8. (a) Representative images and (b) weight loss results of samples exposed in DSW
for four years. Results for Alloy 6063 are not shown. Alloy 6063 exhibited severe pitting
along extrusion lines. Both samples removed after four years exposure had weight loss over
9.5 grams.
-
34
4.1.4. Warm Seawater (WSW) Pre-Treatment
In general, pitting was not severe in the WSW pre-treated samples. Table 4.1-4
summarizes the number of sample scanned per alloy, per exposure time. Although 1-year
samples of Alloy 3003, LA83I, and LA83P samples were removed, they were not scanned with
the profilometer.
Table 4.1-4. Number of scanned warm seawater (WSW) pre-treated samples.
1yr 1.5yr 21mo 2yr 3 year 4 year
1100 3 - - 2 2 2
3003 - - - 2 2 2
5052 2 - - 2 2 2
6063 1 - - 2 2 2
LA83I - - - 2 2 2
LA83P - - - 2 2 2
Of the 4-year samples, Alloys 3003, LA83I, and LA83P had little to no pitting. One
sample of Alloy 5052 had extrusion defects and many small pits were concentrated within the
extrusion defect channel. The other Alloy 5052 sample did not have extrusion defects and no
pits were found. Alloys 1100 and 6063 continue to perform poorly. The maximum pit depth
data for Alloy 1100 and 6063 samples are inaccurate due to incomplete corrosion product
removal. On Alloy 6063, the pits appeared large and shallow. Both samples of Alloy 1100 also
had severe crevice corrosion.
-
35
Figure 4-9. Maximum pit depth and average percent pitted area (per alloy, per exposure time)
for WSW pre-treated samples exposed to CSW.
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
1 yr 2 yr 3 yr 4 yr
Max
imu
m P
it D
epth
[m
m]
Time Exposed
Rack 5 - Maximum Pit Depth
1100 3003 5052 6063 LA83I LA83P
0.0001
0.001
0.01
0.1
1
1 yr 2 yr 3 yr 4 yr
Ave
rage
Per
cen
t P
itte
d A
rea
[%]
Time Exposed
Rack 5 - Average Percent Pitted Area
1100 3003 5052 6063 LA83I LA83P
-
36
(a)
(b)
Figure 4-10. (a) Representative images and (b) weight loss results of samples exposed in
WSW for 40 days, then moved to CSW for four years.
4.2. STEEL TUBE COUPONS
Recent images of the 674-m CSW intake during pipeline inspection showed galvanized
steel hardware (installed during installation of the 674-m CSW pipe at NELHA ~ 30 years ago)
-
37
in good condition, with the galvanized coating still intact. Makai decided to test galvanized steel
as a possible material for the condenser.
Two sets of steel coupons, non-certified and certified have been exposed in CSW. Non-
certified samples were exposed December 2012. Certified samples were exposed to CSW in
June 2013. Coupons were weighed prior to exposure and monitored with electrochemistry. Two
samples of each set were removed in January 2014 to provide 1 year (non-certified) and 6-month
(certified) to provide for weight analysis. Electrochemistry was underestimating the corrosion
rate.
Table 4.2-1. Corrosion rate of certified steel tube samples in CSW
LPR Corrosion Rate [mpy]
Weight Analysis
Corrosion Rate
[mpy]
1-day 3-month 6-month
CS7 0.308 0.208 N.A.
CS8 0.229 0.148 N.A.
CS9 0.136 0.096 2
CS10 0.121 0.084 2.2
Table 4.2-2. Corrosion rate of non-certified steel tube samples in CSW
LPR Corrosion Rate [mpy]
Weight Analysis
Corrosion Rate
[mpy]
28-days 35-days 6-months 9-months 1 year
Upper 0.116 0.177 0.081 0.166 1.9
Lower 0.207 0.176 0.142 0.089 1.8
-
38
Figure 4-11. Certified steel sample after six months in CSW post-removal (top left) and post-
processing (top right). Non-certified steel sample after one year in CSW post-removal (bottom
left) and post-processing (bottom right). Post-processing reveals a relatively smooth wall with
large shallow pits in the certified steel sample and a rough wall with few visible pits in the
non-certified steel sample.
The inner surfaces of both types of tubes were roughened by corrosion product. Some
corrosion product also began to flake off. Post-processing revealed certified steel samples still
had relatively smooth walls that were covered with large shallow pits. The non-certified samples
did not have large shallow pits, but the wall surface was no longer smooth.
-
39
Figure 4-12. Corrosion product beginning to flake on a steel sample tested in the CSW MCIR.
Since the materials cost comprises about 40% of the total heat exchanger cost, a steel
shell-and-tube condenser would provide ~10% cost savings over a similar aluminum condenser.
However, the roughened surfaces will increase headloss (i.e., increased pumping power will be
required) and the flakes of corrosion product will reduce heat transfer. The lowered performance
of a steel condenser would not justify the relatively small cost savings and this test was
discontinued.
4.3. ACID TREATMENT
Acid treatments are being investigated for their ability to clean the metal surface of
impurities and to remove accumulated corrosion product. Nitric acid was first selected for
testing but was found to be too aggressive for aluminum. Nitric acid preferentially attacked
intermetallic compounds, which created pit initiation sites. Currently, sulfamic acid and citric
acid are being tested. Acid concentration, treatment time, and treatment interval are under
investigation. In addition to a set interval, e.g., every two months, Makai is also treating samples
based on observation of OCP.
The acid treatment system consists of a storage/mixing tank and a series of valves and
piping to deliver treatment to a specific sample. The acid treatment is flushed through a systems
testing column for several minutes prior to sample exposure to ensure uniform application. Acid
treatment lines are flushed with freshwater after treatment. Acid treatment is available for
samples in the MCIR. Acid is recirculated through the sample column for two or five minutes
per treatment. For samples with OCP-based intervals, the sample will be treated when OCP rises
above -800 mV. This value is based on previously observed pitting when OCP reaches ~ -700
mV. The tests show that acid treatments can lower OCP to ~ -1000 mV for weeks to months
before the OCP climbs again to ~700 mV.
-
40
Table 4.3-1. Current acid treatment samples in CSW.
Sample Exposure
Date Acid Interval
# of
Treaments
AL 6063-T52 9/13/2013 Citric 2 months 4b
AL 6063-T52 1/9/2014 Citric OCP 3
AL 6063-T52 3/10/2013 Sulfamic 2 months 4a
AL 6063-T52 2/11/2014 Sulfamic OCP 3
AL 6063-T52 12/14/2012 Baseline N/A
aFirst treatment performed immediately before exposure. Removed on 9/24/14.
bFirst treatment performed immediately before exposure. Removed on 11/4/14.
Table 4.3-2. Current acid treatment samples in WSW.
Sample Exposure
Date Acid Interval
# of
Treaments
AL 6063-T52 9/13/2013 Citric 2 months 6a
AL 6063-T52 9/13/2013 Sulfamic 2 months 6a
AL 3003 HT 7/23/2013 Citric 2 months 6b
AL 3003 HT 7/23/2013 Sulfamic 2 months 6b
AL 6063T52 10/19/2012 Baseline N/A
AL 3003 HT 7/23/2013 Baseline N/A
aFirst treatment performed immediately before exposure. Removed on 7/25/2014.
bFirst treatment performed after 2 months of exposure.
Initial trials used 10% (v/v) sulfamic acid with a treatment interval of either every two
months or every six months. In CSW, this concentration may have accelerated pitting, both
samples were removed after 5 months of exposure. The sulfamic treatment concentration has
now been adjusted to 5% (v/v), resulting in a solution with pH = 1.3. Test 46 is treated on a two-
month interval while Test 45 is treated based on OCP.
In Test 46, the sample OCP rose to the steady state value of the baseline sample (~ -720
mV) within 45 days, almost three months sooner than the baseline sample. After the two-month
treatment, OCP dropped to ~ -1000 mV but immediately began to increase. OCP reached and
remained at -720 mV two weeks prior to the four-month treatment. The four-month treatment
dropped OCP to ~ -1000 mV, but OCP returned to -720 mV within 30 days. Corrosion product
-
41
was observed on the trailing edge of the sample after two months. The sample was removed on
9/24/14.
Figure 4-13. Sulfamic acid treatment in CSW. Test 46 2-month treatment dates are indicated
by magenta lines and Test 45 OCP-based treatment dates are indicated by cyan lines. Only the
first 275 days of exposure are shown for the baseline coupon (actual date is on upper x-axis in
green). Treatments have been successful in lowering OCP to ~ -1000 mV. Test 45 was in
stagnant fresh water for several hours due to a flow disruption on 10/1/14.
In Test 45, the OCP has followed the baseline sample, reaching -700 mV after 120 days,
when it was given the first treatment. After the first treatment, OCP remained ~ -1020 mV for
30 days before beginning to increase rapidly. The second treatment was performed 40 days after
the first treatment. After the second treatment, OCP was lowered to -1000 mV but began
increasing immediately. The second treatment lasted for 7 minutes instead of the standard 2
minutes. The third treatment was performed 30 days after the second treatment. OCP after the
third treatment is currently steady at ~ -1020 mV.
Figure 4-14. Before and after sulfamic treatment. Corrosion product buildup was lessened in
this pit, but result was not observed for all pits.
Treatment reduced the
corrosion product in
the pit
Before After
-
42
Some reduction in corrosion product buildup was observed post-treatment for some pits
(Figure 4-14) but not all pits. Removal of corrosion product is believed to prevent the
development of localized environments which could further accelerate pitting.
Citric acid treatments also utilize a 5% (v/v) concentration, which results in a solution
with pH = 2.8.
Figure 4-15. Citric acid treatments in CSW. Test 34 2-month treatment dates are indicated by
magenta lines and Test 38 OCP-based treatment dates are indicated by cyan lines. Only the
first 425 days of exposure are shown for the baseline coupon (actual date is on upper x-axis in
green). Treatments have been successful in lowering OCP to between -900 mV and -1000 mV.
Test 38 was in stagnant fresh water for several hours due to a flow disruption on 10/1/14.
For Test 34, the first treatment was given just prior to starting CSW flow. The two-
month treatment coincides with OCP starting to rise towards pitting potential. Although this is
~60 days faster than the baseline sample, the treatment suppressed OCP to below the baseline
sample. The four-month treatment resulted in an OCP above the pre-treatment value. OCP
continued to rise until it reached OCP of the baseline sample, where it remained until the six-
month treatment (~10 days). After the six-month treatment OCP dropped below the pre-
treatment value and slowly began to increase after 40 days. Pits on the trailing edge of the
sample and some corrosion product buildup were observed at eight months of exposure Figure
4-16). The eight-month treatment dropped OCP but it began to rise immediately. The ten-month
treatment dropped OCP and OCP remained around -1000 mV until the twelve-month treatment.
The sample was removed on 11/4/2014. The pits that were observed at eight months do not
appear to have worsened.
-
43
Figure 4-16. After eight months of exposure, pits and corrosion product buildup were
observed on the trailing edge of the sample receiving citric acid treatment every two months.
For Test 38, OCP rose to the pitting region after 60 days, nearly two months faster than
the baseline sample. The time between each treatments has been different; treatments have been
separated by two months, four months, and two weeks.
In WSW, all samples, including the baseline sample, have OCP below -900 mV (Figure
4-17). Samples were removed on 7/25/14 due to space constraints. No significant observations
on treatment effects could be made.
(a) Alloy 6063-T52 coupons in WSW. Treatment intervals are 2 months; dates are indicated
by purple dashed lines. No ozone treatment was performed on 3/14/2014. Only first 325 days
of exposure are shown for the baseline coupon (actual date is on upper x-axis in green).
Treatment coupons were removed on 7/25/2014 (red line). Baseline coupon is still installed.
-
44
(b) CHART samples (heat treated Alloy 3003). Treatment intervals are 2 months; dates are
indicated by purple dashed lines. No ozone treatment was provided on 3/14/14. OCP
measurements between 6/15/14 – 7/23/14 (pink shaded box) should be ignored due to
malfunctioning reference electrode. Flow disruption on 8/23/14 and 10/1/14 (third and fourth
vertical red lines) resulted in samples being held in stagnant freshwater for several hours.
(c) During sulfamic acid treatment on 11/11/14, the bio-slime layer peeled off the surface of
the CHART coupon, resulting in changes in color and a cleaner coupon surface.
Figure 4-17. Acid and ozone treatments in WSW.
One difference between treatments in WSW vs. CSW lies in the post-treatment OCP. In
WSW, sulfamic acid treatments and ozone drive OCP to ~ -1200 mV. The first citric acid
treatment drove OCP to -1200 mV but subsequent treatments resulted in less negative OCPs.
After the fourth citric acid treatment, OCP only reached -1050 mV, 100 mV less than the steady
state OCP. In CSW, the post-treatment OCP is approximately the same each time.
Reference
electrode
mal-
function
-
45
4.4. OZONE TREATMENT
Makai is comparing the performance of samples given no treatment, one initial treatment,
and multiple treatments on scheduled intervals.
The ozone treatment system is a Water Zone 10 by Ozone Solutions Inc. This unit can
produce 10 gram/hour with its built-in 12 SCFH oxygen concentrator and compressor. Ozone
injection into seawater is accomplished with a 3/4 HP March magnetic drive water recirculation
pump feeding an MK-784 venturi injector. An ambient ozone sensor will signal the controller to
shut the unit down in case of ozone leaks. Currently ozone treatment is available for samples
tested in the MCIR and biofouling rack.
Initial results from the ozone treatment indicate accelerated pitting in ozone treated
samples. Prior studies indicate ozone treatment is beneficial in developing a thicker, denser,
oxide layer, which should provide better corrosion resistance. The appropriate ozone
concentration, treatment time, and treatment interval are under investigation. The current
treatment protocol recirculates ozonated water (1 ppm concentration) for 15 minutes per sample.
In CSW, one sample is being treated every 2 months. Similar to the acid treated samples,
OCP of ozone treated sample reached ~ -700 mV almost 90 days earlier than the baseline sample
(Figure 4-18). Small pits were observed 45 days after exposure. The two-month treatment was
mistakenly missed and the four-month treatment was intentionally skipped to observe pit
development. The six-month treatment was provided on 7/11/2014. The sudden drop in OCP on
6/26/2014 coincides with the removal of a steel sample from the MCIR.
The sulfamic and ozone pre-treatment sample was removed on 8/20/2014. No treatments
were performed after the initial treatment (sulfamic acid at 10% for 2 minutes, ozonated CSW at
1 ppm for 15 min). OCP reached ~ -700 mV 70 days earlier than the baseline sample. Pits were
observed after four months of exposure and by removal, the surface was covered in pits.
-
46
Figure 4-18. OCP of ozone treated samples reached steady-state 70 days (Test 41) and 90 days
(Test 42) before the baseline sample. Ozone treatment dates for Test 42 are indicated by
vertical gray lines. Treatments were planned but not performed on 3/10/2014 and 5/13/2014.
Small pits were observed on 2/25/2014. Only first 275 days of exposure are shown for the
baseline coupon (actual date is on upper x-axis in green). Test 41 was removed on 8/20/2014
and Test 42 was removed on 9/24/14.
In WSW, the samples treated with ozone appeared no different than the acid or baseline
samples (Figure 4-17). Alloy 6063 samples were removed due to space constraints but CHART
samples remain. No significant observations of ozone treatment effects can be made at this time.
Of the acid treatments tested, treatment intervals based on OCP have been more effective
at preventing pitting compared to treatments based on set intervals for samples in CSW. Ozone
treatments have not provided any pitting prevention or protection for the CSW samples.
4.5. CREVICE CORROSION/COATINGS
Previous testing on representative heat exchanger coupons demonstrated extensive
crevice corrosion at the gasket interface. Since these connections are not critical heat transfer
surfaces, suitable coatings could be applied to protect the interface without impacting heat
exchanger performance. Alodine 1201, SiloXel, and 3M 5200 were investigated. Alodine 1201
is a chromic acid based chemical that produces a chrome conversion coating which becomes part
of the aluminum surface. It has been used to improve corrosion resistance in marine
applications, but contains hexavalent chromium which is environmentally hazardous and
requires treatment prior to disposal. SiloXel is a quasi-ceramic conversion coating for aluminum
developed at the University of Hawaii. 3M 5200 is a polyurethane marine adhesive/sealant that
is marketed to resist weathering and salt water.
-
47
Samples of Alloy 3003 and Alloy 6061 were tested along with three coatings to test:
Alodine, SiloXel, and 3M 5200. Two additional coating combinations, 3M 5200 applied to
Alodine or SiloXel coated samples only at the gasket interface, were also tested. Including the
untreated control sample, 6 configurations per alloy were tested.
In general, SiloXel coating integrity was unreliable. Several samples exhibited coating
failures as shown in Figure 4-19.
Figure 4-19. Typical SiloXel coating defect on in-line plate. Samples were still tested.
4.5.1. In-Line Test
The in-line test replicates heat exchanger seawater-side connections. The flow of
seawater is perpendicular to the metal-gasket interface. In-line test samples are tested as a unit
and were exposed to WSW and CSW in February 2013. Performance assessment was limited to
visual inspection of the outside surfaces while the unit was being tested. WSW and CSW units
were opened for interior inspection after one year.
No corrosion was observed in WSW, which is consistent with results from representative
heat exchanger testing. Little information pertaining to coating efficacy could be gathered
without any corroding samples; the WSW test was decommissioned in February 2014.
The CSW unit was decommissioned in March 2014, after 13 months of exposure.
Corrosion was present in all samples except for Alloy 3003 coated with SiloXel and sealed with
polyurethane. Corrosion occurred in the flow channel, starting from the metal-gasket interface,
or on the outside edge of the gasket (Figure 4-20). Coatings provided some degree of protection.
For Alloy 6063, Alodine, Alodine with polyurethane, and SiloXel with polyurethane had at least
one sample that was free of corrosion. For Alloy 3003, all the control and polyurethane coated
samples had corrosion; all other coating combinations protected at least one sample from
corrosion.
-
48
The variability in coating performance may impact the feasibility for use in OTEC heat
exchangers. For example, the discrepancy between the performance of SiloXel with
polyurethane applied to Alloy 3003 versus 6063 may be attributed to defects in the coating
process rather than the performance of the coating itself.
Figure 4-20. In-line sample showing corrosion in the flow channel starting at the metal-
gasket interface and on the outside edge of the gasket.
4.5.2. Cross-Flow Test
The cross-flow test replicates conditions within a plate and frame heat exchanger. The
flow of seawater is parallel to the metal-gasket interface. Cross-flow test samples were tested in
the three-point racks and were exposed to WSW and CSW in January 2013. Photographs were
used to assess sample performance.
Corrosion in the
flow channel starts
at the metal-gasket
interface.
Corrosion
product visible at
the outside edge
of the gasket.
-
49
No corrosion was observed in WSW, which is consistent with results from representative
heat exchanger testing. Little information pertaining to coating efficacy could be gathered
without any corroding samples; the WSW test was decommissioned in February 2014.
The CSW unit was decommissioned in March 2014, after 13 months of exposure.
Results were similar to the in-line test samples. For Alloy 3003, only the control and
polyurethane coated samples were corroded at the metal-gasket interface. The Alodine and
Alodine with polyurethane samples were not corroded at the metal-gasket interface but showed
some corrosion at the edge of the coupons. Only the SiloXel and SiloXel and polyurethane
samples were free of corrosion across the entire sample. For Alloy 6063, the control and
polyurethane coated samples were severely corroded at the metal-gasket interface and pits
covered ~ 60% of the sample surface. Only the SiloXel with polyurethane coating prevented
corrosion at the metal-gasket interface in all the samples. The remaining coating combinations
had one or two samples that still had corrosion.
Similar to the in-line test, the variability in coating performance may impact the
feasibility for use in OTEC heat exchangers.
4.6. BIOFOULING
Biofouling negatively impacts heat exchanger performance by reducing the heat transfer
coefficient and, in severe cases, impeding flow channels. Biofilms can also create local
environments that can accelerate corrosion. Makai has been using hypochlorination for one hour
each day for biofoulant control. Since studies using ozone and iodine treatment have been
promising, Makai decided both warranted further investigation.
The iodine treatment system and protocol was provided by I2 Air Fluid Innovation, Inc.
The system uses an air compressor to force air through cartridges filled with iodine resin. Iodine
vapor is entrained in the air and delivered into the water stream via air bubbles. Iodine treatment
testing was performed on a custom rack.
Iodine infusion testing was inconclusive. Although the testing continued over several
months, equipment malfunctions made it unlikely adequate iodine vapor was entering the
system. The major issues were: 1) seawater backflow into the iodine resin cartridges and 2) light
exposure. Moisture in the cartridges interferes with the entrainment of iodine vapor as air is
forced through the cartridge. Although samples were drawn during infusion, they were sent out
for analysis; the concentration was not determined in real-time and any issues with concentration
were not identified until much later. The equipment complexities could be resolved, but algal
fouling is likely at OTEC intake locations and may travel into the heat exchanger. Algae can
grow even with low level light exposure. Iodine is an anti-bacterial; the biocidal action is not
effective on algae. At this time we will focus on investigating biocides that eliminate algae as
well as bacteria.
-
50
The ozone treatment system is the same as used in pitting mitigation studies. The
original system has been modified to deliver ozonated seawater to the biofouling rack.
Hypochlorination is achieved by using a Chemtech metering pump to pump bleach into
the WSW inlet header for one hour each day. Recent studies in health and water quality
industries emphasize adjusting biocide treatments based on oxidation reduction potential (ORP)
rather than concentration. An ORP around 600 mV provides disinfection whereas an ORP
around 800 mV provides sterilization. A hypochlorite concentration of 550 ppb yields an ORP
of 750 mV. The metering pump has been adjusted to deliver 550 ppb for one hour each day and
ORP will be monitored periodically to determine if further adjustments are required (changes in
water quality could require more or less bleach to achieve the same ORP).
Based on ATP and bacterial plate count results, the current hypochlorination treatment is
effective in controlling biofouling. Although testing in the 1980’s found a chlorine concentration
of 70 ppb was sufficient to control biofouling, our testing found a 250 ppb hypochlorite
concentration did not increase ORP above 400 mV, well below the 600 mV required for
disinfection and the 800 mV required for sterilization. Hypochlorination level is now adjusted to
reach a desired 750 mV in ORP. ORP changes natural fluctuations in water chemistry, but, in
general, a concentration of ~ 550 ppb is required to establish 750 mV.
Ozone treatment may also provide effective biofoulant control, but treatment should be
more frequent than once a month. ATP test results from the biofouling test indicate ozone
provides immediate sterilization (low RLU in the ozone treated samples when swabbed within an
hour of treatment) but fouling begins after the treatment is over (high RLU in ozone treated
samples when swabbed 24 hours after treatment). Furthermore, once a biofilm is established, it
is difficult for any type of treatment to be 100% effective. Bacteria and microbes can use the
film as a protective barrier. One caution to the use of ozone is that in preliminary tests of ozone
treatments in CSW (1 ppm for 5 min, every two months), samples have shown accelerated
localized corrosion. Test duration and sample sizes have been small so it is difficult to draw
conclusions about the effect of ozone on pitting but corrosion must be taken into account when
selecting biofoulant controls.
ATP, bacterial plate count, and photographs of the interior surfaces of the samples in the
biofouling rack confirm the efficacy of hypochlorite treatment at concentrations that raise ORP
to ~ 750 mV for one hour daily.
-
51
5. CONCLUSION AND OUTLOOK
This period’s accomplishments included testing two evaporators and a condenser.
Titanium plate-frame type heat exchangers were tested. The APV condenser had high U values
but also high seawater pressure drop. The APV evaporator had mid-range U values and very low
seawater pressure drop. U-value was limited by ammonia-side convection, which changed very
little with increasing duty. To date, based on corrosion and performance testing results, the
optimum OTEC condenser still remains elusive. Makai is testing prototype versions of a new
heat exchanger design and upcoming efforts will focus on working with manufacturing
techniques to support full-scale, economical production of Makai’s heat exchangers.
Corrosion testing is on-going. The five-year box coupon samples were removed in
November 2014 and will be analyzed shortly. Steel testing has concluded and Makai is no
longer pursuing steel as a potential heat exchanger material. Coating testing has also concluded.
Additional work in coatings has not been planned for heat exchange surfaces but coatings have
potential to protect non-heat transfer surfaces such as tubesheets and manifolds. Acid and ozone
treatment testing will continue. The biofouling study concluded bleach was the most effective
treatment but some accumulation of slime still coats sample surfaces. Modifications to the
bleach system are planned to incorporate monitoring during bleach delivery, finer control of
bleach concentration, and flexibility to treat CSW in addition to WSW. Corrosion efforts will
continue to target pitting and biofouling prevention/mitigation treatments. Makai plans on using
advancements made this year in Makai’s ultra-sonic monitoring capability to provide better
assessment of treatment performance.
A major portion of Makai’s efforts will be focused on turbine operations and
understanding the power generation intricacies of an OTEC system. Turbine installation is
underway and turbine testing is slated for early 2015. The next step would be developing
controls so OTEC-generated electricity could be provided to the grid.
Finally, since Makai’s vision for OTEC’s future is in off-shore facilities, Makai also has
development work planned for cost effective cold water pipe construction and installation
solutions.
OTEC - APRISES11 Task 4.1 OTEC HX Dev and Test_cover pgMakai_OTEC_HTX_2014 Annaul report for APRISES11 final report